wespy-ts 0.2.0

package/README.md

@@ -0,0 +1,146 @@
+# WeSpy (TypeScript)
+
+文章抓取与 Markdown 转换工具 — TypeScript 版本。
+
+## 支持平台
+
+| 平台 | 单篇文章 | 专辑/合集 |
+|------|---------|----------|
+| 微信公众号 (`mp.weixin.qq.com`) | ✅ | ✅ |
+| 掘金 (`juejin.cn`) | ✅ | — |
+| 通用网页 | ✅ | — |
+
+## 安装
+
+### 全局安装
+
+```bash
+npm install -g wespy-ts
+```
+
+### 项目依赖
+
+```bash
+npm install wespy-ts
+```
+
+### 本地开发
+
+```bash
+git clone git@github.com:Cuimc/WeSpy-TS.git
+cd WeSpy-TS
+npm install
+npm run build
+```
+
+## CLI 使用
+
+全局安装后可直接使用 `wespy` 命令，也可以通过 `npx` 运行：
+
+```bash
+# 获取单篇文章（默认输出 Markdown）
+wespy fetch-article https://mp.weixin.qq.com/s/xxxxx
+
+# 使用 npx 运行
+npx wespy fetch-article https://mp.weixin.qq.com/s/xxxxx
+
+# 指定输出格式和目录
+wespy fetch-article https://mp.weixin.qq.com/s/xxxxx -o ./output --all
+
+# 批量下载微信专辑
+wespy fetch-album https://mp.weixin.qq.com/mp/appmsgalbum?... --max 5
+
+# 智能模式（自动判断 URL 类型）
+wespy https://mp.weixin.qq.com/s/xxxxx -o ./output --html --json
+```
+
+### CLI 参数
+
+| 参数 | 说明 | 默认值 |
+|------|------|--------|
+| `-o, --output <dir>` | 输出目录 | `articles` |
+| `--html` | 同时保存 HTML 文件 | false |
+| `--json` | 同时保存 JSON 文件 | false |
+| `--all` | 保存所有格式 | false |
+| `--max <n>` | 专辑最大下载数 | 10 |
+| `--timeout <ms>` | 请求超时时间 | 30000 |
+| `--album-only` | 仅获取文章列表 | false |
+
+## SDK API
+
+```typescript
+import { fetchArticle, fetchAlbum } from 'wespy-ts'
+
+// 获取单篇文章
+const result = await fetchArticle({
+  url: 'https://mp.weixin.qq.com/s/xxxxx',
+  format: ['markdown', 'json'],
+  outputDir: './output',
+})
+
+if (result.ok) {
+  console.log(result.article.title)
+  console.log(result.article.markdown)
+  console.log(result.artifacts)
+} else {
+  console.error(result.error.code, result.error.message)
+}
+
+// 批量获取专辑
+const albumResult = await fetchAlbum({
+  url: 'https://mp.weixin.qq.com/mp/appmsgalbum?...',
+  maxArticles: 10,
+  format: ['markdown'],
+})
+
+if (albumResult.ok) {
+  console.log(`获取了 ${albumResult.articles.length} 篇文章`)
+}
+```
+
+## 输出结构
+
+### ArticleDraft
+
+```typescript
+interface ArticleDraft {
+  platform: 'wechat' | 'juejin' | 'generic'
+  url: string
+  title: string
+  author?: string
+  publishTime?: string
+  contentHtml?: string
+  contentText?: string
+  markdown?: string
+  metadata: Record<string, unknown>
+  fetchedAt: string
+  warnings: string[]
+}
+```
+
+### FetchArticleResult
+
+```typescript
+// 成功
+{ ok: true, article: ArticleDraft, artifacts: OutputArtifact[], warnings: string[] }
+
+// 失败
+{ ok: false, error: { code: string, message: string, retryable: boolean } }
+```
+
+## 测试
+
+```bash
+npm test
+```
+
+## 开发
+
+```bash
+npm run dev    # watch 模式
+npm run lint   # 类型检查
+```
+
+## Python 原版
+
+Python 原版代码保留在 `../WeSpy/` 目录中，未被删除。详见 [MIGRATION.md](./MIGRATION.md)。

package/dist/cli/main.d.ts

@@ -0,0 +1,7 @@
+#!/usr/bin/env node
+/**
+ * WeSpy CLI 入口
+ * 使用 commander 实现，调用 SDK API
+ */
+export {};
+//# sourceMappingURL=main.d.ts.map

package/dist/cli/main.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"main.d.ts","sourceRoot":"","sources":["../../src/cli/main.ts"],"names":[],"mappings":";AACA;;;GAGG"}

package/dist/cli/main.js

@@ -0,0 +1,312 @@
+#!/usr/bin/env node
+/**
+ * WeSpy CLI 入口
+ * 使用 commander 实现，调用 SDK API
+ */
+import { createInterface } from 'node:readline';
+import { readFileSync } from 'node:fs';
+import { join, dirname } from 'node:path';
+import { fileURLToPath } from 'node:url';
+import { Command } from 'commander';
+import { fetchArticle } from '../sdk/fetch-article.js';
+import { fetchAlbum } from '../sdk/fetch-album.js';
+import { fetchAlbumList } from '../sdk/fetch-album-list.js';
+import { isWechatAlbumUrl } from '../platforms/detector.js';
+import { writeJsonSafe, ensureDir } from '../utils/fs.js';
+// 从 package.json 读取版本号，避免硬编码
+const __dirname = dirname(fileURLToPath(import.meta.url));
+const pkg = JSON.parse(readFileSync(join(__dirname, '../../package.json'), 'utf-8'));
+function addCommonOptions(cmd) {
+    return cmd
+        .option('-o, --output <dir>', '输出目录', 'articles')
+        .option('--html', '同时保存 HTML 文件')
+        .option('--json', '同时保存 JSON 文件')
+        .option('--all', '保存所有格式 (HTML + JSON + Markdown)')
+        .option('--download-images', '下载图片到本地（当前迁移版本暂未实现）')
+        .option('--timeout <ms>', '请求超时时间 (毫秒)', '30000')
+        .option('-v, --verbose', '显示详细信息');
+}
+/** 打印文章信息，缺失字段不显示 */
+function printArticleInfo(article) {
+    console.log(`   标题: ${article.title}`);
+    if (article.author)
+        console.log(`   作者: ${article.author}`);
+    if (article.publishTime)
+        console.log(`   时间: ${article.publishTime}`);
+}
+function resolveFormats(options) {
+    if (options.all)
+        return ['markdown', 'html', 'json'];
+    const formats = ['markdown'];
+    if (options.html)
+        formats.push('html');
+    if (options.json)
+        formats.push('json');
+    return formats;
+}
+function printVerbose(options, url, extra) {
+    const formats = resolveFormats(options);
+    console.log('\n[verbose] 配置信息:');
+    console.log(`  URL: ${url}`);
+    console.log(`  输出目录: ${options.output}`);
+    console.log(`  输出格式: HTML=${formats.includes('html')}, JSON=${formats.includes('json')}, Markdown=${formats.includes('markdown')}`);
+    if (extra) {
+        for (const [k, v] of Object.entries(extra)) {
+            console.log(`  ${k}: ${v}`);
+        }
+    }
+}
+// ── 交互模式 ──────────────────────────────────────────────
+async function promptInput(question, defaultVal) {
+    const rl = createInterface({ input: process.stdin, output: process.stdout });
+    return new Promise((resolve) => {
+        rl.question(question, (answer) => {
+            rl.close();
+            resolve(answer.trim() || defaultVal);
+        });
+    });
+}
+async function runInteractiveMode() {
+    console.log('文章获取工具');
+    console.log('='.repeat(40));
+    const url = await promptInput('请输入文章 URL: ', '');
+    if (!url) {
+        console.error('URL 不能为空!');
+        process.exit(1);
+    }
+    const outputDir = await promptInput("输出目录 (回车使用默认 'articles'): ", 'articles');
+    console.log('\n输出格式选择:');
+    console.log('1. 仅 Markdown (默认)');
+    console.log('2. Markdown + HTML');
+    console.log('3. Markdown + JSON');
+    console.log('4. 全部格式 (HTML + JSON + Markdown)');
+    const choice = await promptInput('请选择 (1-4, 回车使用默认1): ', '1');
+    let formats = ['markdown'];
+    if (choice === '2')
+        formats = ['markdown', 'html'];
+    else if (choice === '3')
+        formats = ['markdown', 'json'];
+    else if (choice === '4')
+        formats = ['markdown', 'html', 'json'];
+    const maxArticles = 10;
+    // 执行
+    if (isWechatAlbumUrl(url)) {
+        console.log('\n检测到微信专辑 URL，将批量下载...');
+        const result = await fetchAlbum({
+            url,
+            outputDir,
+            format: formats,
+            maxArticles,
+            timeoutMs: 30000,
+            downloadImages: false,
+        });
+        if (result.ok) {
+            console.log(`\n✅ 批量下载完成! 成功: ${result.articles.length} 篇`);
+        }
+        else {
+            console.error(`\n❌ ${result.error.message}`);
+            process.exit(1);
+        }
+    }
+    else {
+        const result = await fetchArticle({ url, outputDir, format: formats, timeoutMs: 30000, downloadImages: false });
+        if (result.ok) {
+            if ('article' in result) {
+                console.log('\n✅ 成功获取文章!');
+                printArticleInfo(result.article);
+            }
+            else {
+                console.log('\n✅ 批量下载完成!');
+                console.log(`   成功: ${result.articles.length} 篇`);
+                if (result.failedCount)
+                    console.log(`   失败: ${result.failedCount} 篇`);
+            }
+            if (result.artifacts.length > 0) {
+                console.log('   产物:');
+                for (const a of result.artifacts) {
+                    console.log(`     - [${a.type}] ${a.path}`);
+                }
+            }
+        }
+        else {
+            console.error(`\n❌ 获取失败: ${result.error.message}`);
+            process.exit(1);
+        }
+    }
+}
+// ── album-only 处理 ──────────────────────────────────────
+async function handleAlbumOnly(url, options) {
+    const maxArticles = parseInt(options.max, 10);
+    const timeoutMs = parseInt(options.timeout, 10);
+    console.log('仅获取专辑文章列表...');
+    // 使用 fetchAlbumList —— 仅 API 列表，不下载文章（与 Python 一致）
+    const result = await fetchAlbumList({ url, maxArticles, timeoutMs });
+    if (result.ok) {
+        console.log(`\n获取到 ${result.articles.length} 篇文章:`);
+        for (let i = 0; i < result.articles.length; i++) {
+            const a = result.articles[i];
+            console.log(`${String(i + 1).padStart(2)}. ${a.title}`);
+            console.log(`     URL: ${a.url}`);
+            if (a.createTime)
+                console.log(`     时间: ${a.createTime}`);
+            if (i < result.articles.length - 1)
+                console.log();
+        }
+        // 保存文章列表到 JSON（包含完整元数据，与 Python 一致）
+        await ensureDir(options.output);
+        const listFile = join(options.output, `album_articles_${Math.floor(Date.now() / 1000)}.json`);
+        await writeJsonSafe(listFile, result.articles);
+        console.log(`\n文章列表已保存到: ${listFile}`);
+    }
+    else {
+        console.error(`\n❌ 获取失败: ${result.error.message}`);
+        process.exit(1);
+    }
+}
+// ── album 批量下载处理 ────────────────────────────────────
+async function handleAlbumDownload(url, options, formats) {
+    const maxArticles = parseInt(options.max, 10);
+    const timeoutMs = parseInt(options.timeout, 10);
+    console.log('检测到微信专辑 URL，将批量下载...');
+    const result = await fetchAlbum({
+        url,
+        outputDir: options.output,
+        format: formats,
+        maxArticles,
+        timeoutMs,
+        downloadImages: options.downloadImages,
+    });
+    if (result.ok) {
+        const failed = result.failedCount ?? 0;
+        console.log(`\n✅ 批量下载完成!`);
+        console.log(`   成功: ${result.articles.length} 篇`);
+        if (failed > 0)
+            console.log(`   失败: ${failed} 篇`);
+        console.log(`   文章保存在: ${options.output}`);
+        if (result.summaryFile) {
+            console.log(`   汇总: ${result.summaryFile}`);
+        }
+    }
+    else {
+        console.error(`\n❌ 下载失败: ${result.error.message}`);
+        process.exit(1);
+    }
+}
+// ── CLI 程序 ──────────────────────────────────────────────
+const program = new Command();
+program
+    .name('wespy')
+    .description('文章抓取与 Markdown 转换工具')
+    .version(pkg.version);
+// ── fetch-article 子命令 ──────────────────────────────────
+addCommonOptions(program
+    .command('fetch-article')
+    .description('获取单篇文章')
+    .argument('<url>', '文章 URL')).action(async (url, options) => {
+    const formats = resolveFormats(options);
+    const timeoutMs = parseInt(options.timeout, 10);
+    if (options.verbose) {
+        printVerbose(options, url);
+    }
+    const result = await fetchArticle({
+        url,
+        outputDir: options.output,
+        format: formats,
+        timeoutMs,
+        downloadImages: options.downloadImages,
+    });
+    if (result.ok) {
+        if ('article' in result) {
+            console.log('\n✅ 成功获取文章!');
+            printArticleInfo(result.article);
+        }
+        else {
+            console.log('\n✅ 批量下载完成!');
+            console.log(`   成功: ${result.articles.length} 篇`);
+            if (result.failedCount)
+                console.log(`   失败: ${result.failedCount} 篇`);
+        }
+        if (result.artifacts.length > 0) {
+            console.log('   产物:');
+            for (const a of result.artifacts) {
+                console.log(`     - [${a.type}] ${a.path}`);
+            }
+        }
+    }
+    else {
+        console.error(`\n❌ 获取失败: ${result.error.message}`);
+        console.error(`   错误码: ${result.error.code}`);
+        process.exit(1);
+    }
+});
+// ── fetch-album 子命令 ────────────────────────────────────
+addCommonOptions(program
+    .command('fetch-album')
+    .description('批量获取微信专辑文章')
+    .argument('<url>', '专辑 URL')
+    .option('--max <n>', '最大下载文章数', '10')
+    .option('--album-only', '仅获取文章列表，不下载内容')).action(async (url, options) => {
+    if (options.verbose) {
+        printVerbose(options, url, {
+            最大文章数量: parseInt(options.max, 10),
+            仅获取列表: options.albumOnly ?? false,
+        });
+    }
+    if (options.albumOnly) {
+        await handleAlbumOnly(url, options);
+        return;
+    }
+    const formats = resolveFormats(options);
+    await handleAlbumDownload(url, options, formats);
+});
+// ── 默认命令（智能判断 URL 类型）──────────────────────────
+addCommonOptions(program
+    .argument('[url]', '文章或专辑 URL')
+    .option('--max <n>', '专辑最大下载文章数', '10')
+    .option('--album-only', '仅获取文章列表，不下载内容')).action(async (url, options) => {
+    // 无 URL → 交互模式
+    if (!url) {
+        await runInteractiveMode();
+        return;
+    }
+    const formats = resolveFormats(options);
+    const timeoutMs = parseInt(options.timeout, 10);
+    const maxArticles = parseInt(options.max, 10);
+    if (options.verbose) {
+        printVerbose(options, url, { 最大文章数量: maxArticles, 仅获取列表: options.albumOnly ?? false });
+    }
+    if (isWechatAlbumUrl(url)) {
+        if (options.albumOnly) {
+            await handleAlbumOnly(url, options);
+            return;
+        }
+        await handleAlbumDownload(url, options, formats);
+    }
+    else {
+        const result = await fetchArticle({
+            url,
+            outputDir: options.output,
+            format: formats,
+            timeoutMs,
+            downloadImages: options.downloadImages,
+        });
+        if (result.ok) {
+            if ('article' in result) {
+                console.log('\n✅ 成功获取文章!');
+                printArticleInfo(result.article);
+            }
+            else {
+                console.log('\n✅ 批量下载完成!');
+                console.log(`   成功: ${result.articles.length} 篇`);
+                if (result.failedCount)
+                    console.log(`   失败: ${result.failedCount} 篇`);
+            }
+        }
+        else {
+            console.error(`\n❌ ${result.error.message}`);
+            process.exit(1);
+        }
+    }
+});
+program.parse();
+//# sourceMappingURL=main.js.map

package/dist/cli/main.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"main.js","sourceRoot":"","sources":["../../src/cli/main.ts"],"names":[],"mappings":";AACA;;;GAGG;AAEH,OAAO,EAAE,eAAe,EAAE,MAAM,eAAe,CAAA;AAC/C,OAAO,EAAE,YAAY,EAAE,MAAM,SAAS,CAAA;AACtC,OAAO,EAAE,IAAI,EAAE,OAAO,EAAE,MAAM,WAAW,CAAA;AACzC,OAAO,EAAE,aAAa,EAAE,MAAM,UAAU,CAAA;AACxC,OAAO,EAAE,OAAO,EAAE,MAAM,WAAW,CAAA;AACnC,OAAO,EAAE,YAAY,EAAE,MAAM,yBAAyB,CAAA;AACtD,OAAO,EAAE,UAAU,EAAE,MAAM,uBAAuB,CAAA;AAClD,OAAO,EAAE,cAAc,EAAE,MAAM,4BAA4B,CAAA;AAC3D,OAAO,EAAE,gBAAgB,EAAE,MAAM,0BAA0B,CAAA;AAE3D,OAAO,EAAE,aAAa,EAAE,SAAS,EAAE,MAAM,gBAAgB,CAAA;AAEzD,6BAA6B;AAC7B,MAAM,SAAS,GAAG,OAAO,CAAC,aAAa,CAAC,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAA;AACzD,MAAM,GAAG,GAAG,IAAI,CAAC,KAAK,CAAC,YAAY,CAAC,IAAI,CAAC,SAAS,EAAE,oBAAoB,CAAC,EAAE,OAAO,CAAC,CAAwB,CAAA;AAgB3G,SAAS,gBAAgB,CAAC,GAAY;IACpC,OAAO,GAAG;SACP,MAAM,CAAC,oBAAoB,EAAE,MAAM,EAAE,UAAU,CAAC;SAChD,MAAM,CAAC,QAAQ,EAAE,cAAc,CAAC;SAChC,MAAM,CAAC,QAAQ,EAAE,cAAc,CAAC;SAChC,MAAM,CAAC,OAAO,EAAE,iCAAiC,CAAC;SAClD,MAAM,CAAC,mBAAmB,EAAE,qBAAqB,CAAC;SAClD,MAAM,CAAC,gBAAgB,EAAE,aAAa,EAAE,OAAO,CAAC;SAChD,MAAM,CAAC,eAAe,EAAE,QAAQ,CAAC,CAAA;AACtC,CAAC;AAED,qBAAqB;AACrB,SAAS,gBAAgB,CAAC,OAAiE;IACzF,OAAO,CAAC,GAAG,CAAC,UAAU,OAAO,CAAC,KAAK,EAAE,CAAC,CAAA;IACtC,IAAI,OAAO,CAAC,MAAM;QAAE,OAAO,CAAC,GAAG,CAAC,UAAU,OAAO,CAAC,MAAM,EAAE,CAAC,CAAA;IAC3D,IAAI,OAAO,CAAC,WAAW;QAAE,OAAO,CAAC,GAAG,CAAC,UAAU,OAAO,CAAC,WAAW,EAAE,CAAC,CAAA;AACvE,CAAC;AAED,SAAS,cAAc,CAAC,OAAmB;IACzC,IAAI,OAAO,CAAC,GAAG;QAAE,OAAO,CAAC,UAAU,EAAE,MAAM,EAAE,MAAM,CAAC,CAAA;IACpD,MAAM,OAAO,GAAmB,CAAC,UAAU,CAAC,CAAA;IAC5C,IAAI,OAAO,CAAC,IAAI;QAAE,OAAO,CAAC,IAAI,CAAC,MAAM,CAAC,CAAA;IACtC,IAAI,OAAO,CAAC,IAAI;QAAE,OAAO,CAAC,IAAI,CAAC,MAAM,CAAC,CAAA;IACtC,OAAO,OAAO,CAAA;AAChB,CAAC;AAED,SAAS,YAAY,CAAC,OAAmB,EAAE,GAAW,EAAE,KAA+B;IACrF,MAAM,OAAO,GAAG,cAAc,CAAC,OAAO,CAAC,CAAA;IACvC,OAAO,CAAC,GAAG,CAAC,mBAAmB,CAAC,CAAA;IAChC,OAAO,CAAC,GAAG,CAAC,UAAU,GAAG,EAAE,CAAC,CAAA;IAC5B,OAAO,CAAC,GAAG,CAAC,WAAW,OAAO,CAAC,MAAM,EAAE,CAAC,CAAA;IACxC,OAAO,CAAC,GAAG,CAAC,gBAAgB,OAAO,CAAC,QAAQ,CAAC,MAAM,CAAC,UAAU,OAAO,CAAC,QAAQ,CAAC,MAAM,CAAC,cAAc,OAAO,CAAC,QAAQ,CAAC,UAAU,CAAC,EAAE,CAAC,CAAA;IACnI,IAAI,KAAK,EAAE,CAAC;QACV,KAAK,MAAM,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,KAAK,CAAC,EAAE,CAAC;YAC3C,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,KAAK,CAAC,EAAE,CAAC,CAAA;QAC7B,CAAC;IACH,CAAC;AACH,CAAC;AAED,yDAAyD;AAEzD,KAAK,UAAU,WAAW,CAAC,QAAgB,EAAE,UAAkB;IAC7D,MAAM,EAAE,GAAG,eAAe,CAAC,EAAE,KAAK,EAAE,OAAO,CAAC,KAAK,EAAE,MAAM,EAAE,OAAO,CAAC,MAAM,EAAE,CAAC,CAAA;IAC5E,OAAO,IAAI,OAAO,CAAC,CAAC,OAAO,EAAE,EAAE;QAC7B,EAAE,CAAC,QAAQ,CAAC,QAAQ,EAAE,CAAC,MAAM,EAAE,EAAE;YAC/B,EAAE,CAAC,KAAK,EAAE,CAAA;YACV,OAAO,CAAC,MAAM,CAAC,IAAI,EAAE,IAAI,UAAU,CAAC,CAAA;QACtC,CAAC,CAAC,CAAA;IACJ,CAAC,CAAC,CAAA;AACJ,CAAC;AAED,KAAK,UAAU,kBAAkB;IAC/B,OAAO,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAA;IACrB,OAAO,CAAC,GAAG,CAAC,GAAG,CAAC,MAAM,CAAC,EAAE,CAAC,CAAC,CAAA;IAE3B,MAAM,GAAG,GAAG,MAAM,WAAW,CAAC,aAAa,EAAE,EAAE,CAAC,CAAA;IAChD,IAAI,CAAC,GAAG,EAAE,CAAC;QACT,OAAO,CAAC,KAAK,CAAC,WAAW,CAAC,CAAA;QAC1B,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAA;IACjB,CAAC;IAED,MAAM,SAAS,GAAG,MAAM,WAAW,CAAC,4BAA4B,EAAE,UAAU,CAAC,CAAA;IAE7E,OAAO,CAAC,GAAG,CAAC,WAAW,CAAC,CAAA;IACxB,OAAO,CAAC,GAAG,CAAC,oBAAoB,CAAC,CAAA;IACjC,OAAO,CAAC,GAAG,CAAC,oBAAoB,CAAC,CAAA;IACjC,OAAO,CAAC,GAAG,CAAC,oBAAoB,CAAC,CAAA;IACjC,OAAO,CAAC,GAAG,CAAC,kCAAkC,CAAC,CAAA;IAE/C,MAAM,MAAM,GAAG,MAAM,WAAW,CAAC,sBAAsB,EAAE,GAAG,CAAC,CAAA;IAE7D,IAAI,OAAO,GAAmB,CAAC,UAAU,CAAC,CAAA;IAC1C,IAAI,MAAM,KAAK,GAAG;QAAE,OAAO,GAAG,CAAC,UAAU,EAAE,MAAM,CAAC,CAAA;SAC7C,IAAI,MAAM,KAAK,GAAG;QAAE,OAAO,GAAG,CAAC,UAAU,EAAE,MAAM,CAAC,CAAA;SAClD,IAAI,MAAM,KAAK,GAAG;QAAE,OAAO,GAAG,CAAC,UAAU,EAAE,MAAM,EAAE,MAAM,CAAC,CAAA;IAE/D,MAAM,WAAW,GAAG,EAAE,CAAA;IAEtB,KAAK;IACL,IAAI,gBAAgB,CAAC,GAAG,CAAC,EAAE,CAAC;QAC1B,OAAO,CAAC,GAAG,CAAC,wBAAwB,CAAC,CAAA;QACrC,MAAM,MAAM,GAAG,MAAM,UAAU,CAAC;YAC9B,GAAG;YACH,SAAS;YACT,MAAM,EAAE,OAAO;YACf,WAAW;YACX,SAAS,EAAE,KAAK;YAChB,cAAc,EAAE,KAAK;SACtB,CAAC,CAAA;QACF,IAAI,MAAM,CAAC,EAAE,EAAE,CAAC;YACd,OAAO,CAAC,GAAG,CAAC,mBAAmB,MAAM,CAAC,QAAQ,CAAC,MAAM,IAAI,CAAC,CAAA;QAC5D,CAAC;aAAM,CAAC;YACN,OAAO,CAAC,KAAK,CAAC,OAAO,MAAM,CAAC,KAAK,CAAC,OAAO,EAAE,CAAC,CAAA;YAC5C,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAA;QACjB,CAAC;IACH,CAAC;SAAM,CAAC;QACN,MAAM,MAAM,GAAG,MAAM,YAAY,CAAC,EAAE,GAAG,EAAE,SAAS,EAAE,MAAM,EAAE,OAAO,EAAE,SAAS,EAAE,KAAK,EAAE,cAAc,EAAE,KAAK,EAAE,CAAC,CAAA;QAC/G,IAAI,MAAM,CAAC,EAAE,EAAE,CAAC;YACd,IAAI,SAAS,IAAI,MAAM,EAAE,CAAC;gBACxB,OAAO,CAAC,GAAG,CAAC,aAAa,CAAC,CAAA;gBAC1B,gBAAgB,CAAC,MAAM,CAAC,OAAO,CAAC,CAAA;YAClC,CAAC;iBAAM,CAAC;gBACN,OAAO,CAAC,GAAG,CAAC,aAAa,CAAC,CAAA;gBAC1B,OAAO,CAAC,GAAG,CAAC,UAAU,MAAM,CAAC,QAAQ,CAAC,MAAM,IAAI,CAAC,CAAA;gBACjD,IAAI,MAAM,CAAC,WAAW;oBAAE,OAAO,CAAC,GAAG,CAAC,UAAU,MAAM,CAAC,WAAW,IAAI,CAAC,CAAA;YACvE,CAAC;YACD,IAAI,MAAM,CAAC,SAAS,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;gBAChC,OAAO,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAA;gBACrB,KAAK,MAAM,CAAC,IAAI,MAAM,CAAC,SAAS,EAAE,CAAC;oBACjC,OAAO,CAAC,GAAG,CAAC,WAAW,CAAC,CAAC,IAAI,KAAK,CAAC,CAAC,IAAI,EAAE,CAAC,CAAA;gBAC7C,CAAC;YACH,CAAC;QACH,CAAC;aAAM,CAAC;YACN,OAAO,CAAC,KAAK,CAAC,aAAa,MAAM,CAAC,KAAK,CAAC,OAAO,EAAE,CAAC,CAAA;YAClD,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAA;QACjB,CAAC;IACH,CAAC;AACH,CAAC;AAED,0DAA0D;AAE1D,KAAK,UAAU,eAAe,CAAC,GAAW,EAAE,OAAmB;IAC7D,MAAM,WAAW,GAAG,QAAQ,CAAC,OAAO,CAAC,GAAG,EAAE,EAAE,CAAC,CAAA;IAC7C,MAAM,SAAS,GAAG,QAAQ,CAAC,OAAO,CAAC,OAAO,EAAE,EAAE,CAAC,CAAA;IAE/C,OAAO,CAAC,GAAG,CAAC,cAAc,CAAC,CAAA;IAE3B,mDAAmD;IACnD,MAAM,MAAM,GAAG,MAAM,cAAc,CAAC,EAAE,GAAG,EAAE,WAAW,EAAE,SAAS,EAAE,CAAC,CAAA;IAEpE,IAAI,MAAM,CAAC,EAAE,EAAE,CAAC;QACd,OAAO,CAAC,GAAG,CAAC,SAAS,MAAM,CAAC,QAAQ,CAAC,MAAM,OAAO,CAAC,CAAA;QACnD,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,MAAM,CAAC,QAAQ,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;YAChD,MAAM,CAAC,GAAG,MAAM,CAAC,QAAQ,CAAC,CAAC,CAAE,CAAA;YAC7B,OAAO,CAAC,GAAG,CAAC,GAAG,MAAM,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,KAAK,EAAE,CAAC,CAAA;YACvD,OAAO,CAAC,GAAG,CAAC,aAAa,CAAC,CAAC,GAAG,EAAE,CAAC,CAAA;YACjC,IAAI,CAAC,CAAC,UAAU;gBAAE,OAAO,CAAC,GAAG,CAAC,YAAY,CAAC,CAAC,UAAU,EAAE,CAAC,CAAA;YACzD,IAAI,CAAC,GAAG,MAAM,CAAC,QAAQ,CAAC,MAAM,GAAG,CAAC;gBAAE,OAAO,CAAC,GAAG,EAAE,CAAA;QACnD,CAAC;QAED,oCAAoC;QACpC,MAAM,SAAS,CAAC,OAAO,CAAC,MAAM,CAAC,CAAA;QAC/B,MAAM,QAAQ,GAAG,IAAI,CAAC,OAAO,CAAC,MAAM,EAAE,kBAAkB,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,GAAG,EAAE,GAAG,IAAI,CAAC,OAAO,CAAC,CAAA;QAC7F,MAAM,aAAa,CAAC,QAAQ,EAAE,MAAM,CAAC,QAAQ,CAAC,CAAA;QAC9C,OAAO,CAAC,GAAG,CAAC,eAAe,QAAQ,EAAE,CAAC,CAAA;IACxC,CAAC;SAAM,CAAC;QACN,OAAO,CAAC,KAAK,CAAC,aAAa,MAAM,CAAC,KAAK,CAAC,OAAO,EAAE,CAAC,CAAA;QAClD,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAA;IACjB,CAAC;AACH,CAAC;AAED,uDAAuD;AAEvD,KAAK,UAAU,mBAAmB,CAAC,GAAW,EAAE,OAAmB,EAAE,OAAuB;IAC1F,MAAM,WAAW,GAAG,QAAQ,CAAC,OAAO,CAAC,GAAG,EAAE,EAAE,CAAC,CAAA;IAC7C,MAAM,SAAS,GAAG,QAAQ,CAAC,OAAO,CAAC,OAAO,EAAE,EAAE,CAAC,CAAA;IAE/C,OAAO,CAAC,GAAG,CAAC,sBAAsB,CAAC,CAAA;IACnC,MAAM,MAAM,GAAG,MAAM,UAAU,CAAC;QAC9B,GAAG;QACH,SAAS,EAAE,OAAO,CAAC,MAAM;QACzB,MAAM,EAAE,OAAO;QACf,WAAW;QACX,SAAS;QACT,cAAc,EAAE,OAAO,CAAC,cAAc;KACvC,CAAC,CAAA;IAEF,IAAI,MAAM,CAAC,EAAE,EAAE,CAAC;QACd,MAAM,MAAM,GAAG,MAAM,CAAC,WAAW,IAAI,CAAC,CAAA;QACtC,OAAO,CAAC,GAAG,CAAC,aAAa,CAAC,CAAA;QAC1B,OAAO,CAAC,GAAG,CAAC,UAAU,MAAM,CAAC,QAAQ,CAAC,MAAM,IAAI,CAAC,CAAA;QACjD,IAAI,MAAM,GAAG,CAAC;YAAE,OAAO,CAAC,GAAG,CAAC,UAAU,MAAM,IAAI,CAAC,CAAA;QACjD,OAAO,CAAC,GAAG,CAAC,aAAa,OAAO,CAAC,MAAM,EAAE,CAAC,CAAA;QAC1C,IAAI,MAAM,CAAC,WAAW,EAAE,CAAC;YACvB,OAAO,CAAC,GAAG,CAAC,UAAU,MAAM,CAAC,WAAW,EAAE,CAAC,CAAA;QAC7C,CAAC;IACH,CAAC;SAAM,CAAC;QACN,OAAO,CAAC,KAAK,CAAC,aAAa,MAAM,CAAC,KAAK,CAAC,OAAO,EAAE,CAAC,CAAA;QAClD,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAA;IACjB,CAAC;AACH,CAAC;AAED,2DAA2D;AAE3D,MAAM,OAAO,GAAG,IAAI,OAAO,EAAE,CAAA;AAE7B,OAAO;KACJ,IAAI,CAAC,OAAO,CAAC;KACb,WAAW,CAAC,qBAAqB,CAAC;KAClC,OAAO,CAAC,GAAG,CAAC,OAAO,CAAC,CAAA;AAEvB,0DAA0D;AAE1D,gBAAgB,CACd,OAAO;KACJ,OAAO,CAAC,eAAe,CAAC;KACxB,WAAW,CAAC,QAAQ,CAAC;KACrB,QAAQ,CAAC,OAAO,EAAE,QAAQ,CAAC,CAC/B,CAAC,MAAM,CAAC,KAAK,EAAE,GAAW,EAAE,OAAmB,EAAE,EAAE;IAClD,MAAM,OAAO,GAAG,cAAc,CAAC,OAAO,CAAC,CAAA;IACvC,MAAM,SAAS,GAAG,QAAQ,CAAC,OAAO,CAAC,OAAO,EAAE,EAAE,CAAC,CAAA;IAE/C,IAAI,OAAO,CAAC,OAAO,EAAE,CAAC;QACpB,YAAY,CAAC,OAAO,EAAE,GAAG,CAAC,CAAA;IAC5B,CAAC;IAED,MAAM,MAAM,GAAG,MAAM,YAAY,CAAC;QAChC,GAAG;QACH,SAAS,EAAE,OAAO,CAAC,MAAM;QACzB,MAAM,EAAE,OAAO;QACf,SAAS;QACT,cAAc,EAAE,OAAO,CAAC,cAAc;KACvC,CAAC,CAAA;IAEF,IAAI,MAAM,CAAC,EAAE,EAAE,CAAC;QACd,IAAI,SAAS,IAAI,MAAM,EAAE,CAAC;YACxB,OAAO,CAAC,GAAG,CAAC,aAAa,CAAC,CAAA;YAC1B,gBAAgB,CAAC,MAAM,CAAC,OAAO,CAAC,CAAA;QAClC,CAAC;aAAM,CAAC;YACN,OAAO,CAAC,GAAG,CAAC,aAAa,CAAC,CAAA;YAC1B,OAAO,CAAC,GAAG,CAAC,UAAU,MAAM,CAAC,QAAQ,CAAC,MAAM,IAAI,CAAC,CAAA;YACjD,IAAI,MAAM,CAAC,WAAW;gBAAE,OAAO,CAAC,GAAG,CAAC,UAAU,MAAM,CAAC,WAAW,IAAI,CAAC,CAAA;QACvE,CAAC;QACD,IAAI,MAAM,CAAC,SAAS,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YAChC,OAAO,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAA;YACrB,KAAK,MAAM,CAAC,IAAI,MAAM,CAAC,SAAS,EAAE,CAAC;gBACjC,OAAO,CAAC,GAAG,CAAC,WAAW,CAAC,CAAC,IAAI,KAAK,CAAC,CAAC,IAAI,EAAE,CAAC,CAAA;YAC7C,CAAC;QACH,CAAC;IACH,CAAC;SAAM,CAAC;QACN,OAAO,CAAC,KAAK,CAAC,aAAa,MAAM,CAAC,KAAK,CAAC,OAAO,EAAE,CAAC,CAAA;QAClD,OAAO,CAAC,KAAK,CAAC,WAAW,MAAM,CAAC,KAAK,CAAC,IAAI,EAAE,CAAC,CAAA;QAC7C,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAA;IACjB,CAAC;AACH,CAAC,CAAC,CAAA;AAEF,0DAA0D;AAE1D,gBAAgB,CACd,OAAO;KACJ,OAAO,CAAC,aAAa,CAAC;KACtB,WAAW,CAAC,YAAY,CAAC;KACzB,QAAQ,CAAC,OAAO,EAAE,QAAQ,CAAC;KAC3B,MAAM,CAAC,WAAW,EAAE,SAAS,EAAE,IAAI,CAAC;KACpC,MAAM,CAAC,cAAc,EAAE,eAAe,CAAC,CAC3C,CAAC,MAAM,CAAC,KAAK,EAAE,GAAW,EAAE,OAAmB,EAAE,EAAE;IAClD,IAAI,OAAO,CAAC,OAAO,EAAE,CAAC;QACpB,YAAY,CAAC,OAAO,EAAE,GAAG,EAAE;YACzB,MAAM,EAAE,QAAQ,CAAC,OAAO,CAAC,GAAG,EAAE,EAAE,CAAC;YACjC,KAAK,EAAE,OAAO,CAAC,SAAS,IAAI,KAAK;SAClC,CAAC,CAAA;IACJ,CAAC;IAED,IAAI,OAAO,CAAC,SAAS,EAAE,CAAC;QACtB,MAAM,eAAe,CAAC,GAAG,EAAE,OAAO,CAAC,CAAA;QACnC,OAAM;IACR,CAAC;IAED,MAAM,OAAO,GAAG,cAAc,CAAC,OAAO,CAAC,CAAA;IACvC,MAAM,mBAAmB,CAAC,GAAG,EAAE,OAAO,EAAE,OAAO,CAAC,CAAA;AAClD,CAAC,CAAC,CAAA;AAEF,iDAAiD;AAEjD,gBAAgB,CACd,OAAO;KACJ,QAAQ,CAAC,OAAO,EAAE,WAAW,CAAC;KAC9B,MAAM,CAAC,WAAW,EAAE,WAAW,EAAE,IAAI,CAAC;KACtC,MAAM,CAAC,cAAc,EAAE,eAAe,CAAC,CAC3C,CAAC,MAAM,CAAC,KAAK,EAAE,GAAuB,EAAE,OAAmB,EAAE,EAAE;IAC9D,eAAe;IACf,IAAI,CAAC,GAAG,EAAE,CAAC;QACT,MAAM,kBAAkB,EAAE,CAAA;QAC1B,OAAM;IACR,CAAC;IAED,MAAM,OAAO,GAAG,cAAc,CAAC,OAAO,CAAC,CAAA;IACvC,MAAM,SAAS,GAAG,QAAQ,CAAC,OAAO,CAAC,OAAO,EAAE,EAAE,CAAC,CAAA;IAC/C,MAAM,WAAW,GAAG,QAAQ,CAAC,OAAO,CAAC,GAAG,EAAE,EAAE,CAAC,CAAA;IAE7C,IAAI,OAAO,CAAC,OAAO,EAAE,CAAC;QACpB,YAAY,CAAC,OAAO,EAAE,GAAG,EAAE,EAAE,MAAM,EAAE,WAAW,EAAE,KAAK,EAAE,OAAO,CAAC,SAAS,IAAI,KAAK,EAAE,CAAC,CAAA;IACxF,CAAC;IAED,IAAI,gBAAgB,CAAC,GAAG,CAAC,EAAE,CAAC;QAC1B,IAAI,OAAO,CAAC,SAAS,EAAE,CAAC;YACtB,MAAM,eAAe,CAAC,GAAG,EAAE,OAAO,CAAC,CAAA;YACnC,OAAM;QACR,CAAC;QACD,MAAM,mBAAmB,CAAC,GAAG,EAAE,OAAO,EAAE,OAAO,CAAC,CAAA;IAClD,CAAC;SAAM,CAAC;QACN,MAAM,MAAM,GAAG,MAAM,YAAY,CAAC;YAChC,GAAG;YACH,SAAS,EAAE,OAAO,CAAC,MAAM;YACzB,MAAM,EAAE,OAAO;YACf,SAAS;YACT,cAAc,EAAE,OAAO,CAAC,cAAc;SACvC,CAAC,CAAA;QAEF,IAAI,MAAM,CAAC,EAAE,EAAE,CAAC;YACd,IAAI,SAAS,IAAI,MAAM,EAAE,CAAC;gBACxB,OAAO,CAAC,GAAG,CAAC,aAAa,CAAC,CAAA;gBAC1B,gBAAgB,CAAC,MAAM,CAAC,OAAO,CAAC,CAAA;YAClC,CAAC;iBAAM,CAAC;gBACN,OAAO,CAAC,GAAG,CAAC,aAAa,CAAC,CAAA;gBAC1B,OAAO,CAAC,GAAG,CAAC,UAAU,MAAM,CAAC,QAAQ,CAAC,MAAM,IAAI,CAAC,CAAA;gBACjD,IAAI,MAAM,CAAC,WAAW;oBAAE,OAAO,CAAC,GAAG,CAAC,UAAU,MAAM,CAAC,WAAW,IAAI,CAAC,CAAA;YACvE,CAAC;QACH,CAAC;aAAM,CAAC;YACN,OAAO,CAAC,KAAK,CAAC,OAAO,MAAM,CAAC,KAAK,CAAC,OAAO,EAAE,CAAC,CAAA;YAC5C,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAA;QACjB,CAAC;IACH,CAAC;AACH,CAAC,CAAC,CAAA;AAEF,OAAO,CAAC,KAAK,EAAE,CAAA"}

package/dist/converter/html-to-markdown.d.ts

@@ -0,0 +1,9 @@
+/**
+ * HTML → Markdown 转换器
+ * 基于 cheerio 解析 + 递归遍历，忠实还原 Python 版行为
+ */
+/**
+ * 将 HTML 字符串转换为 Markdown
+ */
+export declare function htmlToMarkdown(html: string): string;
+//# sourceMappingURL=html-to-markdown.d.ts.map

package/dist/converter/html-to-markdown.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"html-to-markdown.d.ts","sourceRoot":"","sources":["../../src/converter/html-to-markdown.ts"],"names":[],"mappings":"AAAA;;;GAGG;AAOH;;GAEG;AACH,wBAAgB,cAAc,CAAC,IAAI,EAAE,MAAM,GAAG,MAAM,CAInD"}

package/dist/converter/html-to-markdown.js

@@ -0,0 +1,171 @@
+/**
+ * HTML → Markdown 转换器
+ * 基于 cheerio 解析 + 递归遍历，忠实还原 Python 版行为
+ */
+import * as cheerio from 'cheerio';
+import { buildProxyImageUrl } from '../utils/url.js';
+import { detectCodeLanguage, cleanCodeContent } from '../utils/text.js';
+/**
+ * 将 HTML 字符串转换为 Markdown
+ */
+export function htmlToMarkdown(html) {
+    if (!html)
+        return '';
+    const $ = cheerio.load(html, { xmlMode: false });
+    return convertNode($, $.root());
+}
+function convertNode($, node) {
+    let md = '';
+    node.contents().each((_, child) => {
+        if (child.type === 'text') {
+            const text = $(child).text().trim();
+            if (text)
+                md += text;
+            return;
+        }
+        if (child.type !== 'tag')
+            return;
+        const el = $(child);
+        const tag = child.tagName?.toLowerCase();
+        switch (tag) {
+            case 'br':
+                md += '\n';
+                break;
+            case 'p':
+            case 'div':
+            case 'section': {
+                const content = convertNode($, el).trim();
+                if (content)
+                    md += '\n\n' + content + '\n';
+                break;
+            }
+            case 'h1':
+            case 'h2':
+            case 'h3':
+            case 'h4':
+            case 'h5':
+            case 'h6': {
+                const level = parseInt(tag[1], 10);
+                const content = convertNode($, el).trim();
+                if (content)
+                    md += '\n' + '#'.repeat(level) + ' ' + content + '\n';
+                break;
+            }
+            case 'strong':
+            case 'b': {
+                const content = convertNode($, el).trim();
+                if (content)
+                    md += '**' + content + '**';
+                break;
+            }
+            case 'em':
+            case 'i': {
+                const content = convertNode($, el).trim();
+                if (content)
+                    md += '*' + content + '*';
+                break;
+            }
+            case 'img': {
+                const src = el.attr('data-src') ?? el.attr('src') ?? '';
+                const alt = el.attr('alt') ?? '';
+                if (src) {
+                    const proxySrc = buildProxyImageUrl(src);
+                    md += `\n![${alt}](${proxySrc})\n`;
+                }
+                break;
+            }
+            case 'a': {
+                const href = el.attr('href') ?? '';
+                const text = convertNode($, el).trim();
+                if (href && text) {
+                    md += `[${text}](${href})`;
+                }
+                else if (text) {
+                    md += text;
+                }
+                break;
+            }
+            case 'ul':
+            case 'ol': {
+                const listContent = convertList($, el, tag === 'ol');
+                if (listContent)
+                    md += '\n' + listContent + '\n';
+                break;
+            }
+            case 'code': {
+                // 行内 code（如果父元素是 pre，由 pre 处理）
+                const parent = child.parent;
+                if (parent && parent.type === 'tag' && parent.tagName?.toLowerCase() === 'pre')
+                    break;
+                const codeContent = el.text().trim();
+                if (codeContent)
+                    md += '`' + codeContent + '`';
+                break;
+            }
+            case 'pre': {
+                const codeContent = extractCodeFromPre(el);
+                const language = detectPreLanguage($, el);
+                if (codeContent) {
+                    if (language) {
+                        md += `\n\`\`\`${language}\n${codeContent}\n\`\`\`\n`;
+                    }
+                    else {
+                        md += `\n\`\`\`\n${codeContent}\n\`\`\`\n`;
+                    }
+                }
+                break;
+            }
+            default: {
+                md += convertNode($, el);
+            }
+        }
+    });
+    return md;
+}
+function convertList($, listEl, ordered) {
+    let md = '';
+    let index = 1;
+    listEl.children('li').each((_, item) => {
+        const content = convertNode($, $(item)).trim();
+        if (content) {
+            if (ordered) {
+                md += `${index}. ${content}\n`;
+                index++;
+            }
+            else {
+                md += `- ${content}\n`;
+            }
+        }
+    });
+    return md;
+}
+function extractCodeFromPre(preEl) {
+    const codeEl = preEl.find('code');
+    const raw = codeEl.length > 0 ? codeEl.text() : preEl.text();
+    return cleanCodeContent(raw);
+}
+function detectPreLanguage(_$, preEl) {
+    // 从 pre 的 class 检测
+    const preClasses = (preEl.attr('class') ?? '').split(/\s+/).filter(Boolean);
+    let lang = detectCodeLanguage(preClasses);
+    if (lang)
+        return lang;
+    // 从内部 code 的 class 检测
+    const codeEl = preEl.find('code');
+    if (codeEl.length > 0) {
+        const codeClasses = (codeEl.attr('class') ?? '').split(/\s+/).filter(Boolean);
+        lang = detectCodeLanguage(codeClasses);
+        if (lang)
+            return lang;
+        // data-language 属性
+        const dataLang = codeEl.attr('data-language') ?? codeEl.attr('lang');
+        if (dataLang)
+            return dataLang.toLowerCase();
+    }
+    // pre 的 data-language 属性
+    const preDataLang = preEl.attr('data-language') ?? preEl.attr('lang');
+    if (preDataLang)
+        return preDataLang.toLowerCase();
+    return null;
+}
+//# sourceMappingURL=html-to-markdown.js.map